Increased sensitivity to thyroid hormone in mice with
complete deficiency of thyroid hormone receptor
␣
Paolo E. Macchia*, Yoko Takeuchi†, Tomoko Kawai†, Kevin Cua*, Karine Gauthier‡, Olivier Chassande‡, Hisao Seo†, Yoshitaka Hayashi†, Jacques Samarut‡, Yoshiharu Murata†, Roy E. Weiss*, and Samuel Refetoff*§¶储
Departments of *Medicine and§Pediatrics and the¶J. P. Kennedy Jr. Mental Retardation Research Center, University of Chicago, MC 3090, 5841 South Maryland Avenue, Chicago, IL 60637;†Research Institute of Environmental Medicine, University of Nagoya,
Nagoya 464-8601, Japan; and‡Centre National de la Recherche Scientifique, Unite´ Mixte de Recherche 49, Institut National de la Recherche Agronomique, LA 913, Ecole Normale Supe´rieure, 69364 Lyon, France
Edited by J. E. Rall, National Institutes of Health, Bethesda, MD, and approved October 31, 2000 (received for review July 3, 2000)
Only three of the four thyroid hormone receptor (TR) isoforms,␣1, 1,
and2, bind thyroid hormone (TH) and are considered to be true TRs.
TR␣2 binds to TH response elements on DNA, but its role in vivo is still
unknown. We produced mice completely deficient in TR␣ (TR␣o/o)
that maintain normal serum thyroid-stimulating hormone (TSH) con-centration despite low serum thyroxine (T4), suggesting increased
sensitivity to TH. We therefore examined the effects of TH (L-3,3
ⴕ,5-triiodothyronine, L-T3) given to TH-deprived and to intact TR␣o/omice.
Controls were wild-type (WT) mice of the same strain and mice resistant to TH due to deficiency in TR (TRⴚ/ⴚ). In liver, T3produced
significantly greater responses in TR␣o/oand smaller responses in
TRⴚ/ⴚas compared with WT mice. In contrast, cardiac responses to
L-T3were absent or reduced in TR␣o/o, whereas they were similar in
WT and TRⴚ/ⴚmice, supporting the notion that TR␣1 is the
domi-nant TH-dependent TR isoform in heart. 5-Triiodothyronine (L-T3)
given to intact mice produced a greater suppression of serum T4in
TR␣o/othan it did in WT mice and reduced by a greater amount the
TSH response to TSH-releasing hormone. This is an in vivo demon-stration that a TR deficiency can enhance sensitivity to TH. This effect is likely due to the abrogation of the constitutive ‘‘silencing’’ effect of TR␣2 in tissues expressing the TR isoforms.
T
hyroid hormone (TH) action is mediated through specific nuclear TH receptors (TRs) functioning as ligand-dependent transcription factors that increase or decrease the expression of target genes (1–3). There are two TR genes. The TR locus generates the1 and 2 receptors by using two different promoters and alternative splicing. The TR␣ locus encodes TR␣1, a molecule containing 410 aa, and three proteins that do not bind the ligand, triiodothyronine (T3) (4–7) (GenBank accession no. AI322342).TR␣2, which results from an alternative splicing of the TR␣ primary transcript, has 492 aa, the first 370 of which are identical to TR␣1. TR⌬␣1 and TR⌬␣2, molecules of approximately 154 and 237 aa, respectively, are generated from an internal promoter located in intron 7 of the TR␣ locus (6). The function of TR␣2, which has a wide tissue distribution (8), and that of TR⌬␣1 and TR⌬␣2, found mainly in brain, gut, and lung (6), are still unknown. However, all have been shown to inhibit ligand-dependent trans-activation of TR and TR␣1, a phenomenon termed the dominant negative effect (6, 9–11,**).
The relative contribution of the two TR gene products in mediating TH responses is poorly understood because of the paucity of information regarding in vivo function. In vitro DNA binding studies and functional assays in transfected cells have generated conflicting results concerning the specific effect of TR isoforms in gene regulation (12, 13). Interpretation is compli-cated, however, because data are derived from artificial systems using the overexpression of chimeric gene constructs that may not be faithful models of events occurring in the intact animal. The use of knockout mice that lack a particular TR isoform provides a powerful tool for exploring the relative contribution of each of the TR isoforms to the TH-mediated regulation of various biological processes in different tissues. The specific role of TR␣ was investigated indirectly in mice deficient in TR (14).
Results suggested that TR␣ plays a major role in the regulation of heart rate and basal metabolism (15). This conclusion was confirmed and attributed to the TR␣1 isoform (15).
Herein we examined the effects of TH deprivation and treatment withL-3,3⬘,5-triiodothyronine (L-T3) on the pituitary,
liver, and heart of homozygous mice deficient in all forms of TR␣ (TR␣o/o). Compared with those of wild-type (WT) mice, the
pituitary gland and liver of TR␣o/omice were more sensitive to
TH, whereas the same tissues of TR knockout (TR⫺/⫺) mice,
as previously shown, were resistant to TH. In contrast, the heart of TR␣o/omice was hyporesponsive to TH, indicating that TR␣
is the predominant TR form expressed in this tissue. Given that TR␣1 knockout mice have no increased sensitivity to TH (16), our results suggest that TR␣2 exerts a dominant negative effect in some tissues in vivo and that the absence of TR␣2 enhances sensitivity to TH.
Experimental Methods
Animal Procedures.As is the case with the TR␣⫺/⫺mouse (17), the
TR␣o/omouse is deficient in both TR␣1 and TR␣2 encoded by
the TR␣ gene, but, in addition, it lacks the shorter transcripts initiated from the internal promoter located in intron 7 (6). However, the gene sequence for Rev-erbA␣, the TR-related protein encoded by the opposite strand of TR␣ (18, 19), remains intact. In contrast to the TR␣⫺/⫺mice, the TR␣o/omice survive
without rescue by TH treatment. They are fertile but have a slightly slower growth rate. Thus the dose of L-T3was adjusted
to the body weight (BW). Mice resistant to TH due to deletion of the TR gene (TR⫺/⫺) (described in ref. 20) were used for
comparison. The TR⫺/⫺mice were derived after nine crossings
of TR⫺/⫹heterozygote with WT mice of the C57BL兾6J strain
before inbreeding. TR␣o/⫹ mice were crossed four times with
WT mice of the same strain before heterozygotes were inter-crossed. Breeding of the resulting homozygotes TR␣o/o and
TR␣⫹/⫹(WT) mice was used in the grouped experiments.
Mice were weaned in the fourth week after birth and were fed rodent diet containing 0.53 ppm iodine (no. 5053; LabDiet, Brentwood, MO) ad libitum and given tap water. All animal
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: TH, thyroid hormone; TR, TH receptor; TH, thyroid hormone; TSH, thyroid-stimulating hormone; T4, thyroxine; L-T3, L-3,3⬘,5-triiodothyronine; BW, body weight; TRH, TSH-releasing hormone; 5⬘DI, iodothyronine type I 5⬘-deiodinase; ME, malic enzyme; SERCA2, sarcoplasmic reticulum calcium adenosine triphosphatase; PTU, 5-propyl-2-thio-uracil; WT, wild-type.
储To whom reprint requests should be addressed. E-mail: refetoff@medicine.uchicago. bsd.edu.
**This paper was presented in part at the 82nd Annual Meeting of the Endocrine Society, June 21–24, 2000, Toronto, Canada, and selected as a finalist for the Knoll Award. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Article published online before print: Proc. Natl. Acad. Sci. USA, 10.1073兾pnas.011306998. Article and publication date are at www.pnas.org兾cgi兾doi兾10.1073兾pnas.011306998
experiments were performed at the University of Chicago according to protocols approved by the Institutional Animal Care and Use Committee.
Because of sex and age differences in serum thyroid-stimulating hormone (TSH) and thyroxine (T4) concentrations,
all mice were male and were 60–70 days old at the beginning of each experiment. The BW of TR␣o/o mice (mean⫾ SEM) of
18.4⫾ 0.4 g was significantly lower (P ⬍ 0.0001) than that of the WT (23.4⫾ 0.5 g) and TR⫺/⫺mice (22.0⫾ 0.4 g). Experiments
were terminated by exsanguination under methoxyflurane (Pitman–Moore, Mundelein, IL) anesthesia. Livers were com-pletely removed, weighed, cut into 0.2–0.5-g pieces, and frozen on dry ice before storage at⫺85°C. Atria were removed and heart ventricles were immediately frozen and stored at⫺85°C.
Induction of Hypothyroidism and Treatment with TH.TH deficiency
was induced in 10–14 male mice of each type (TR␣o/o, WT, and
TR⫺/⫺) with a low-iodine diet containing 0.15%
5-propyl-2-thiouracil (PTU), as described (21). On the 11th day, animals of each type were split into two groups. One group received daily i.p. injections of 2g of L-T3兾100 g BW兾day for 4 days, whereas
the other group received only the vehicle. The PTU diet was given throughout the L-T3and vehicle treatment period. The
concentration of L-T3 was confirmed by radioimmunoassay.
Experiments were terminated 12–16 h after the last injection. In a separate experiment, groups of 4–5 mice were similarly treated and blood was collected for serum T3determination 24 h after
the penultimate L-T3dose and at 2, 4, 8, and 16 h after the last
L-T3injection. Note that separate animals were used for the
2-and 8-h 2-and 4- 2-and 16-h posttreatment blood sampling. In a separate experiment, L-T3was given on the same schedule
to groups of mice of each type (10 WT and 7 TR␣o/o) with no prior
induction of hypothyroidism. Two incremental doses were given consecutively as described (22), 0.2g兾100 g BW兾day, followed by 0.8g兾100 g BW兾day. Blood samples (approximately 300 l) were obtained from the tail vein before the treatment (baseline) and 12–16 h after the last injection of each incremental L-T3dose.
The suppressive effect of TH on the response of TSH to TSH-releasing hormone (TRH) was examined in WT and TR␣o/o
mice. Three groups of 5–6 animals of each type were given the two doses of L-T3or the vehicle only (basal) for 4 days. On the fifth day,
12–16 h after the last injection, blood samples were obtained before and 15 min after the i.p. administration of 0.275g TRH.
Measurements in Serum Samples. Cholesterol levels were
mea-sured with an autoanalyzer as described. Serum TSH was measured with a sensitive radioimmunoassay (23), and results were expressed in bioassayable TSH units. Serum total T4and T3
concentrations were measured by radioimmunoassays (Diagnos-tic Products, Los Angeles). The sensitivities of these assays were 0.2g T4兾dl and 20 ng T3兾dl.
Heart Rate.Electrocardiograms were recorded under chloral
hy-drate anesthesia (4 mg兾10 g BW i.p.) with mice placed in a water bed maintained at 37°-38°C. A Hewlett–Packard monitor兾terminal (model 78534AA) with a chart speed of 25 mm兾s was used.
Tissue Analyses. T3 binding to liver nuclei was determined as
described (15). Pools of approximately 1 g of frozen liver pieces from at least four animals of each type were used. All animals had received the PTU diet for 14 days and were, thus, hypothy-roid before liver collection. T3 binding affinity and maximal
binding capacity were determined by the method of Scatchard. The former was corrected for DNA content.
Northern analyses were performed on RNA extracted from tissues of individual animals. Total RNA was prepared by the acid–guanidinium thiocyanate—phenol–chloroform method (24) from livers and heart ventricles frozen 1–3 min after death and kept at⫺85°C. Fifteen micrograms of total RNA was denatured and fractionated by electrophoresis on 0.8% agarose gel, transferred onto GeneScreen Plus (DuPont兾NEN), using VacuGene (Amer-sham Pharmacia). cDNA probes for iodothyronine type I 5⬘-deiodinase (5⬘DI) (25), malic enzyme (ME) (26), and sarcoplasmic reticulum calcium adenosine triphosphatase (SERCA2) [prepared by reverse transcription–PCR with the sense (5⬘-ACGATCTGT-GCTCTGTGTAATGACTCT-3⬘) and antisense (5⬘-GCGCG-TCGTTCACACCATCACCAGTCA-3⬘) oligonucleotides] were labeled with [␥-32P]deoxycytidine triphosphate (specific activity 111
Bq兾mmol) (DuPont兾NEN), using the Random Primed DNA La-beling Kit (Roche Molecular Biochemicals). Hybridization and subsequent washing procedures were carried out as described (27). The quantity of mRNA on radiographs was measured with a molecular imager (Bio-Rad).
To assess the uniformity in RNA transfer among samples, hybridized membranes were reprobed with a cDNA comple-mentary to rRNA. To correct the results of Northern blots, the activity of rRNA probe hybridized to each lane was divided by the average activity of rRNA for the entire group. These ratios were used to correct the activity of the respective mRNA abundance in each lane.
Data Presentation and Statistical Analysis.Values are reported as
mean ⫾ SE. P values were calculated by two-way ANOVA or Student’s t test when comparisons were made within the same genotype. Values corresponding to the respective limits of the assay sensitivities were assigned to samples that measured below the detectable range. Outliers were identified by a two-tailed test with a significance level of⬍0.05 (28, 29). No more than one outlier per treatment group was removed. One TR⫺/⫺mouse did not achieve
a sufficient degree of hypothyroidism with the PTU diet, as determined by a serum TSH level of 725 milliunits兾liter and a T4
of 5.3g兾dl, both 10-fold below and above the group means. All data generated from this mouse were excluded. Statistics were
Table 1. Thyroid function tests and heart rate in adult male TR␣o兾omice as compared with WT
and Trⴚ兾ⴚcontrols TSH, mU兾liter T4, g兾dl T3, ng兾dl Heart rate, beats per min TR␣o兾o 22⫾ 3.5 3.29⫾ 0.09 96.1⫾ 7.2 358⫾ 22 (32) (31) (11) (23) WT 25⫾ 3.0 3.76⫾ 0.10 83.6⫾ 3.0 470⫾ 6 (54) (47) (13) (43) TR⫺兾⫺ 136⫾ 18 7.78⫾ 0.33 142.3⫾ 12.4 522⫾ 8 (42) (37) (12) (48) P (TR␣o兾ovs. WT) NS ⬍0.005 NS ⬍0.0001 P (TR␣o兾ovs. TR⫺兾⫺) ⬍0.0001 ⬍0.001 ⬍0.005 ⬍0.0001 The number of animals is in parentheses. NS, not significant.
performed usingSTATVIEW 5.0software (Abacus Concepts,
Berke-ley, CA). Results
Thyroid Function Tests and Heart Rate at Baseline in TR␣o/oMice as
Compared with WT and TRⴚ/ⴚMice.Results of thyroid function
tests in untreated TR␣o/oare compared with those in WT and
TR⫺/⫺mice (Table 1). Despite the similar levels of circulating
TSH, TR␣o/o mice have significantly lower levels of total T4
(3.29⫾ 0.09 g兾dl) compared with WT animals (3.76 ⫾ 0.10 g兾dl, P ⬍ 0.005). As previously reported (14, 21), the TR⫺/⫺
animals were resistant to TH and had increased levels of serum T4and T3as well as TSH concentration. The mean basal heart
rate of TR␣o/omice was significantly lower than that of WT mice,
a finding similar to that reported for mice lacking only the TR␣1 isoform (16). The higher heart rate in TR⫺/⫺mice is as reported
previously (15, 30) and is consistent with the higher serum TH levels in these animals.
Sensitivity of Serum TSH to the Feedback Regulation by TH. The
observation of a small, but significant, decrease in serum T4
concentration not associated with an alteration in basal serum TSH values could be explained by reduced T4binding to serum proteins,
mild hypothyroidism, or increased sensitivity to TH. To determine the physiological significance of this observation, groups of mice were given two incremental doses of L-T3, and the suppressive
effect on serum T4was examined (Fig. 1). In addition, because the
basal serum TSH concentrations were too low for accurate mea-surement of a decrease induced by L-T3, the TSH response 15 min
after the administration of TRH was determined (Fig. 2). Treatment with 0.2g L-T3兾100 g BW brought the serum T4
of TR␣o/oanimals to 2.26⫾ 0.01g兾dl as compared with 3.01 ⫾
0.11g兾dl in WT animals. This represents a reduction of 34.0 ⫾ 5.4% and 18.2⫾ 3.7%, respectively (P ⬍ 0.05). Treatment with the higher dose of 0.8g L-T3兾100 g BW兾day for 4 more days
produced the same degree of serum T4suppression in the two
types of mice (Fig. 1).
The response of TSH to TRH before TH treatment and its suppression after the administration of the same two doses of L-T3 are shown in Fig. 2. In animals not given L-T3, the
incremental TSH response to TRH was not different in TR␣o/o
mice compared with WT mice but was significantly lower in TR⫺/⫺mice (P⬍ 0.01). The higher dose of L-T3suppressed the
TSH response by 77.3⫾ 5.7% in the TR␣o/omice as compared
with 42.8⫾ 15.8% in WT mice (P ⫽ 0.02). Collectively, these
data indicate that the feedback regulation of the pituitary– thyroid axis of TR␣o/omice is more sensitive to TH.
Of interest is the lower response of TSH to TRH at baseline in TR⫺/⫺mice and the 49.2⫾ 26.3% paradoxical response in
these mice given the lower dose of L-T3. Similar findings were
reported 20 years ago with a comparable dose of L-T3(31) given
to humans homozygous for TR gene deletion (32).
Thyroid Function in TH-Deprived and TH-Treated Mice.To investigate
the role of TR␣ in the mediation of TH action, mice were studied during TH deprivation and after supplementation with TH. The dose of 2g L-T3兾100 g BW (0.5g兾25 g mouse) has been shown
to be effective in bringing the high serum TSH of hypothyroid WT mice to normal, inducing metabolic effects, and producing changes in liver (15, 21, 33). This investigative approach served two purposes. First, it allowed determination of the tissue responses of TR␣o/omice to the same amount of TH as in WT
and TR⫺/⫺mice, which, under basal conditions, have different
serum concentrations of TH. Second, observations under the conditions of TH deprivation and full replacement provided a measure of the maximal physiological effect of TH.
The three genotypes that received the PTU diet only had similar serum TSH, T4, and T3levels (Table 2). L-T3treatment
produced a greater suppression of serum TSH in the TR␣o/omice
than in the WT animals, a difference that did not reach statistical significance. As expected, the effect on TR⫺/⫺mice was lesser,
reflecting their decreased sensitivity to TH (Table 2).
Serum T3concentrations achieved at different times after the Fig. 1. Effect of TH on the feedback suppression of thyroid function. TR␣o/o
and WT mice were treated with two incremental doses of L-T3each for 4 consecutive days. The decrease in serum T4was expressed as a percentage of the baseline values measured in the same mice before L-T3treatment. Note the greater suppression of T4in TR␣o/omice receiving the lower L-T3dose. The numbers of animals are in parentheses.
Fig. 2. Effect of TH on the TSH response to TRH. Groups of mice from the three genotypes were treated for 4 days with two different doses of L-T3or with the vehicle only. TSH was measured before and 15 min after the admin-istration of TRH. (A) TSH response to TRH in mice not given L-T3(basal response). (B) TSH responses to TRH in mice given L-T3, expressed as a per-centage of the basal response. Note the paradoxical increase in TR⫺/⫺mice receiving 0.2g L-T3兾100 g BW and the greater suppressive effect of 0.8g L-T3兾100 g BW in TR␣o/omice. In A, the number of animals in each group is in parentheses. In B, data for each bar are derived from five or six animals. P values indicated by*are differences in responses after L-T3treatment as compared with baseline. P values above bars were obtained by two-way ANOVA of original numerical data.
administration of L-T3were not different among the animals of
the three genotypes (Table 2 and Fig. 3).
Role of TR␣ in the Mediation of TH Action on Liver.Four known
markers of TH action were measured during TH deprivation and after its supplementation as described above. L-T3 treatment
reduced the liver weight of WT and TR␣o/omice but, as expected
(33), had no effect on the TR⫺/⫺ animals. However, the
decrease was greater in TR␣o/o mice than in WT mice (P ⬍
0.005) (Fig. 4). TR␣o/omice had a significantly higher reduction
in cholesterol after treatment with L-T3as compared with WT
animals (P ⬍ 0.005). In contrast, and as expected (15), the T3-mediated reduction in serum cholesterol was significantly
smaller in TR⫺/⫺mice (P⬍ 0.0001) (Fig. 3).
5⬘DI and ME are well characterized TH-responsive genes, the transcriptional regulations of which have been extensively studied in the rat (3, 34) and have been shown to undergo similar changes in response to L-T3in the mouse (15, 33). cDNAs of these two genes
were used to determine their mRNA content in the liver of TR␣o/o,
WT, and TR⫺/⫺mice by Northern blotting (Fig. 5). TH
depriva-tion abolished the ability to detect 5⬘DI mRNA, even after pro-longed exposure of the hybridized membranes. L-T3 treatment
produced a significantly greater increase in the abundance of the mRNA of this enzyme in TR␣o/o mice as compared with WT
animals and a much smaller increase in TR⫺/⫺mice (P⬍ 0.005
and⬍ 0.0001, respectively). ME mRNA was also significantly (P ⬍ 0.001) higher in TR␣o/othan in WT mice under L-T3treatment.
Quantification of T3Binding TR in Liver.To determine whether the absence of TR␣ produces a compensatory increase in TR, T3
binding studies were carried out in liver nuclei. The maximal T3
binding capacity in three separate experiments averaged 144
fmol兾100g DNA for livers of WT animals. The corresponding values in TR␣o/oand TR⫺/⫺mice were, respectively, 66% and
21% of the WT. These results indicate that deficiency in one of the TR genes does not produce a compensatory increase in the other. The relative contributions of TR and TR␣1 in the liver of mice are similar to those reported in the rat (35).
Responses to TH of Heart Rate and TH-Responsive Genes Expressed in Heart.Compared with baseline, TH deprivation decreased the
heart rate of both WT and TR⫺/⫺mice but had no effect on
TR␣o/omice (358⫾ 22 vs. 335 ⫾ 20 beats兾min). L-T3treatment
produced an increase in heart rate in WT and TR⫺/⫺ mice
(89⫾ 8% and 66 ⫾ 4%) that was significantly higher in relative magnitude than that in TR␣o/omice (27⫾ 7%) (Fig. 6A).
Relative to the state of TH deprivation, L-T3 produced a
modest increase in cardiac SERCA2 mRNA in the WT and TR⫺/⫺mice but not in the TR␣o/omice (P⬍ 0.001) (Fig. 6B). Discussion
TH exerts its effects through interaction with specific nuclear TRs,␣1, 1, and 2 (36), that regulate the expression of target genes by binding to cognate TH response elements, either as homodimers or as heterodimers in association with retinoid X receptors (37, 38). Because both TR␣ and TR genes are expressed in most tissues (8), either one or both can be impli-cated in mediating in vivo the TH-dependent effects.
Fig. 3. Serum T3concentrations at different times after the administration of L-T3. Blood was obtained at the indicated times after the fourth i.p. dose of L-T3on the last day of the experiment. Values are not significantly different among the three genotypes at each time point.
Fig. 4. Effect TH deprivation and L-T3treatment on liver weight and serum cholesterol. Groups of mice from the three genotypes were TH-deprived (PTU) and then treated with TH (PTU⫹ T3). Shown are measured values (Left) and the percent changes induced by L-T3treatment (Right). Note the greater relative response to TH of TR␣o/omice.
Table 2. Thyroid function tests in TR␣o兾o, WT and TRⴚ兾ⴚmice after treatment with PTU and PTUⴙT3
PTU* PTU⫹ T3† TSH, mU兾liter T4,g兾dl T3, ng兾dl TSH, mU兾L T4,g兾dl T3, ng兾dl TR␣o兾o 10,650⫾ 1,900 0.71⫾ 0.08 50⫾ 8 8⫾ 1.5 0.37⫾ 0.04 147⫾ 24 WT 8,700⫾ 1,150 1.04⫾ 0.07 48⫾ 4 18⫾ 7 0.38⫾ 0.05 114⫾ 23 TR⫺兾⫺ 10,300⫾ 1,850 0.93⫾ 0.11 34⫾ 11 274⫾ 83 0.95⫾ 0.08 142⫾ 46 P (TR␣o兾ovs. WT) NS NS NS NS NS NS P (TR␣o兾ovs. TR⫺兾⫺) NS NS NS ⬍0.01 ⬍0.0001 NS NS, not significant.
*Fifteenth day of treatment with PTU.
†At the termination of the experiment, after 4 days of L-T
In our work, we sought to determine the role of TR␣ by deleting all of its isoforms and examining the resulting tissue responses to TH. Data were compared with those observed in WT mice as well as animals deficient in both TR isoforms, manifesting signs of resistance to TH (20). The targeted muta-tion in the TR␣o/omouse inactivates completely not only the two
full-length TR␣ gene transcripts encoded by the main promoter, but also the two shorter transcripts, TR⌬␣1 and TR⌬␣2, derived from the promoter located in intron 7 (6). However, the gene sequence for Rev-erbA␣, the TR-related protein encoded by the opposite strand of TR␣ (18, 19), remains intact.
Similar to the TR␣1-deficient mice reported by Wikstro¨m et
al. (16), TR␣o/omice, of the same age and sex, had significantly
lower serum T4levels and heart rates than the respective WT
animals. However, in contrast to the TR␣1-deficient mice, the serum TSH concentrations of TR␣o/o mice were not different
from those of the WT animals, even after the analysis of more than 30 serum samples of each genotype. The maintenance of a normal serum TSH concentration despite a lower concentration of T4suggested increased sensitivity to TH. Indeed,
adminis-tration of the same dose of L-T3to TR␣o/oand WT mice had a
greater suppressive effect on the serum T4levels of the former,
indicating a greater inhibition of basal TSH, even though the corresponding lower serum TSH values did not reach statistical significance. Evidence for a reduced secretion of TSH in the TR␣o/o mice was secured from the greater L-T3-induced
sup-pression of the TSH response to TRH. Together, these results indicate that the feedback regulation of the pituitary–thyroid axis of TR␣o/omice is more sensitive to TH.
We then examined the sensitivity to TH in two peripheral tissues, liver and heart, that express predominantly the TR and TR␣ genes, respectively (15, 39–40, ††). Studies were carried out under the conditions of TH deprivation and TH supplementa-tion, according to a previously tested and standardized protocol (15, 21, 33). This protocol allows the comparison of tissue responses to the same amount of TH in mice that under basal conditions have different concentrations of TH. Moreover, the comparison of measurements obtained under the conditions of TH deprivation and full replacement increases the ability to detect modest effects of the hormone. It should be noted that this method does not achieve complete TH deficiency. However, serum TSH and T4concentrations were not significantly
differ-ent among mice of the three genotypes. Furthermore, similar levels of serum T3were achieved in all genotypes during L-T3
treatment.
Again TR␣o/omice showed evidence for increased sensitivity
to TH in yet another tissue, the liver. L-T3given to TH-deprived
animals reduced significantly more the liver weight and serum cholesterol concentration in TR␣o/oas compared with WT mice.
The former reflects glycogen loss through TH-mediated catab-olism and the latter, TH-induced stimulated transcription of low-density lipoprotein receptor and cholesterol 7-␣ hydroxylase genes, leading to enhanced removal of low-density lipoprotein and cholesterol (41, 42). L-T3also produced a greater increase ††Kinugawa, K., Minobe, W. A., Wood, W., Ridgway, C. E., Lowes, B. D., Baxter, J. D., Riberio,
R. C. J., Long, C. S. & Bristow, M. R. (1999) Circulation 100, Suppl. 18, I–507 (abstr.). Fig. 5. Effect of TH deprivation and L-T3 treat-ment on the expression of TH-regulated genes in liver. Mice from the three genotypes were TH-deprived (PTU) and then treated with TH (PTU⫹ L-T3). Liver RNAs were analyzed for their abun-dance in specific mRNAs by Northern blotting. Radioautographs of membranes hybridized with the two labeled cDNA probes (only five consecu-tive animals of each group) are shown on the left, and the percent changes induced by L-T3are plot-ted on the right. The membrane with RNA from TR␣o/oand the corresponding WT mice was ex-posed for 3 days (¶); those with RNA from the TR⫺/⫺and another set of RNAs from the same WT mice were exposed for 4 days. The same mem-branes were rehybridized with an 18S rRNA probe. Note the greater relative response to TH of TR␣o/omice.
Fig. 6. Effect of TH deprivation and L-T3treatment on heart rate and the expression of a TH-regulated gene in heart ventricle. Mice from the three genotypes were TH-deprived (PTU) and then treated with TH (PTU⫹ L-T3). (A) Mean heart rates (Left) and the percentage changes induced by the treatment with L-T3(Right). (B) The radioautographs of membranes hybridized with the SERCA2 probe (five consecutive animals of each group) are on the left, and the percentage changes induced by L-T3 are plotted on the right. The same membranes were rehybridized with an 18S rRNA probe. Note the reduced responses of SRECA2 to L-T3in TR␣o/omice.
in the two TH-inducible liver mRNAs, 5⬘DI and ME (26, 34, 43, 44), in TR␣o/othan in WT animals. It is of interest to note that
under the current experimental conditions, TH deprivation reduced the expression of 5⬘DI mRNA to undetectable levels in animals of the three genotypes, whereas the reduction in ME mRNA was less pronounced. As is the case with the up-regulation of the pituitary TSH gene (20, 21, 45), down-regulation of the liver 5⬘DI gene, during TH deprivation, does not appear to require the presence of TR.
TR␣o/omice have a significant reduction in heart rate, which
increased with the administration of L-T3, as previously reported
for TR␣1 knockout mice (16). Of interest is our finding that, in contrast to the WT and TR⫺/⫺mice, the heart rate of TR␣o/o
mice did not change significantly during TH deprivation. This observation suggests that in absence of TR␣ the heart rate is in the hypothyroid range but still can be increased by TH, probably through the residual TR. Nevertheless, this L-T3-induced
in-crease in the TR␣o/omice was smaller in magnitude compared
with the WT and TR⫺/⫺mice. The effect of L-T3on SERCA2
mRNA was abolished in TR␣o/omice. Judging from the results
obtained with the TR⫺/⫺mice, the TH-mediated response of
SERCA2 is completely TR␣-dependent.
The observations made in the TR⫺/⫺mice are in agreement
with those previously reported (15, 22) for TR⫺/⫺ mice
pro-duced by a slightly different method of gene disruption (14). A finding not previously reported in the TR knockout mouse is the significantly lower response of TSH to TRH at baseline and the paradoxical increase of this response after the administra-tions of the lower L-T3 dose. Although the reason for this
unexpected response cannot be explained, this same effect has been reported in humans genotypically and phenotypically sim-ilar to the TR⫺/⫺mice (31, 32) but not in humans with TH
resistance caused by a dominant negative mutant TR (46). Increased sensitivity of TR␣o/omice to TH was demonstrated
in the pituitary and liver, both tissues expressing predominantly the TR isoform. These mice differ from the TR␣1-deficient mice that have persistent expression of TR␣2 and, presumably, TR⌬␣2. Compensatory increases in TR expression do not
appear to occur at the RNA and protein levels in brain and liver of TR␣1-deficient and TR␣o/o mice, respectively. Thus, the
increased sensitivity of TR␣o/omice to TH cannot be explained
by a higher level of TR gene expression as observed in livers of mice overexpressing TR (33). It is thus likely that the enhanced responsiveness to TH in the TR␣o/o mouse is due to the
abrogation of the constitutive silencing mediated through TR␣2. This ligand-unresponsive TR has been shown to exert a weak dominant negative effect in transfected cells (10, 11). The ubiquitous expression of TR␣2 and its presence in tissues with predominantly TR-mediated TH action are expected to atten-uate the effect of TH, particularly when TH is present in small amounts. The enhanced effect of TH should not be manifested in tissues of TR␣o/o mice in which TH action is mediated
predominantly through TR␣1, as is the case in the heart. It should be noted that our data do not exclude the possible contribution of TR␣1 deficiency to the increased sensitivity to TH, by removal of a weaker transactivator that normally com-petes with the TR. The loss of the TR⌬␣ isoforms has no effect on the observed effects on liver and pituitary because they are not expressed in these tissues‡‡. Proof that a deficiency in TR␣2
is the basis for the observed hypersensitivity to TH in mice completely deficient of TR␣ should be contained in the mouse deficient in TR␣2 only. Although such mice do exist (47), no information is yet available regarding their precise phenotype. We thank the following investigators for the provision of cDNA probes: 5⬘DI, P. Reed Larsen; ME, Vera M. Nikodem; and SERCA2, Takashi Nagaya. This work was supported in part by Grant DK 15,070 from the National Institutes of Health to S.R.; the Seymor J. Abrams Thyroid Research Center; a grant from the Ministry of Health and Welfare, Japan, to H.S.; and Grants-in-Aid for Scientific Research (B) 11472225 to Y.M. and 10470226 to H.S. from the Ministry of Education, Science, Sports and Culture, Japan.
1. Lazar, M. A. (1993) Endocr. Rev. 14, 184–193. 2. Chin, W. W. (1994) Thyroid 4, 389–393.
3. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Schutz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., et al. (1995) Cell 83, 835–839. 4. Prost, E., Koenig, R. J., Moore, D. D., Larsen, P. R. & Whalen, R. G. (1988) Nucleic Acids
Res. 16, 6248.
5. Moeller, M., Rapoport, B. & Gavin, L. A. (1989) J. Neuroendocrinol. 5, 351–356. 6. Chassande, O., Fraichard, A., Gauthier, K., Flamant, F., Legrand, C., Savatier, P., Laudet,
V. & Samarut, J. (1997) Mol. Endocrinol. 11, 1278–1290.
7. Laudet, V., Begue, A., Henry-Duthoit, C., Joubel, A., Martin, P., Stehelin, D. & Saule, S. (1991) Nucleic Acids Res. 19, 1105–1112.
8. Hodin, R. A., Lazar, M. A. & Chin, W. W. (1990) J. Clin. Invest. 85, 101–105. 9. Lazar, M. A., Hodin, R. A. & Chin, W. W. (1989) Proc. Natl. Acad. Sci. USA 86, 7771–7774. 10. Burgos-Trinidad, M. & Koenig, R. J. (1999) Mol. Cell. Endocrinol. 149, 107–114. 11. Tagami, T., Kopp, P., Johnson, W., Arseven, O. K. & Jameson, J. L. (1998) Endocrinology
139,2535–2544.
12. Hollenberg, A. N., Monden, T., Flynn, T. R., Boers, M. E., Cohen, O. & Wondisford, F. E. (1995) Mol. Endocrinol. 9, 540–550.
13. Feng, P., Li, Q. L., Satoh, T. & Wilber, J. F. (1994) Biochem. Biophys. Res. Commun. 200, 171–177.
14. Forrest, D., Hanebuth, E., Smeyne, R. J., Everds, N., Stewart, C. L., Wehner, J. M. & Curran, T. (1996) EMBO J. 15, 3006–3015.
15. Weiss, R. E., Murata, Y., Cua, K., Hayashi, Y., Seo, H. & Refetoff, S. (1998) Endocrinology 139,4945–4952.
16. Wikstro¨m, L., Johansson, C., Salto, C., Barlow, C., Campos Barros, A., Baas, F., Forrest, D., Thoren, P. & Vennstrom, B. (1998) EMBO J. 17, 455–461.
17. Fraichard, A., Chassande, O., Plateroti, M., Roux, J. P., Trouillas, J., Dehay, C., Legrand, C., Gauthier, K., Kedinger, M., Malaval, L., et al. (1997) EMBO J. 16, 4412–4420. 18. Lazar, M. A., Hodin, R. A., Darling, D. S. & Chin, W. W. (1989) Mol. Cell. Biol. 9,
1128–1136.
19. Spanjaard, R. A., Nguyen, V. P. & Chin, W. W. (1994) Mol. Endocrinol. 8, 286–295. 20. Gauthier, K., Chassande, O., Plateroti, M., Roux, J. P., Legrand, C., Pain, B., Rousset, B.,
Weiss, R., Trouillas, J. & Samarut, J. (1999) EMBO J. 18, 623–631.
21. Weiss, R. E., Forrest, D., Pohlenz, J., Cua, K., Curran, T. & Refetoff, S. (1997) Endocrinology 138,3624–3629.
22. Weiss, R. E., Xu, J., Ning, G., Pohlenz, J., O’Malley, B. W. & Refetoff, S. (1999) EMBO J. 18,1900–1904.
23. Pohlenz, J., Maqueem, A., Cua, K., Weiss, R. E., Van Sande, J. & Refetoff, S. (1999) Thyroid 9,1265–1271.
24. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156–159. 25. Berry, M. J., Banu, L. & Larsen, P. R. (1991) Nature (London) 349, 438–440. 26. Magnuson, M. A., Morioka, H., Tecce, M. F. & Nikodem, V. M. (1986) J. Biol. Chem. 261,
1183–1186.
27. Murata, Y., Seo, H., Sekiguchi, K., Imai, T., Lee, J. & Matsui, N. (1990) Mol. Endocrinol. 4,693–699.
28. Grubbs, F. (1969) Technometrics 11, 1–21.
29. Grubbs, F. E. & Beck, G. (1972) Technometrics 14, 847–854.
30. Johansson, C., Go¨the, S., Forrest, D., Vennsrto¨m, B. & Thore´n, P. (1999) Am. J. Physiol. 276, H2006–H2012.
31. Refetoff, S., Degroot, L. J. & Barsano, C. P. (1980) J. Clin. Endocrinol. Metab. 51, 41–45. 32. Takeda, K., Sakurai, A., DeGroot, L. J. & Refetoff, S. (1992) J. Clin. Endocrinol. Metab. 74,
49–55.
33. Hayashi, Y., Mangoura, D. & Refetoff, S. (1996) Mol. Endocrinol. 10, 100–106. 34. Menjo, M., Murata, Y., Fujii, T., Nimura, Y. & Seo, H. (1993) Endocrinology 133,
2984–2990.
35. Schwartz, H., Lazar, M. & Oppenheimer, J. (1994) J. Biol. Chem. 269, 24777–24782. 36. Lazar, M. A. & Chin, W. W. (1990) J. Clin. Invest. 86, 1777–1782.
37. Yu, V. C., Delsert, C., Andersen, B., Holloway, J. M., Devary, O. V., Naar, A. M., Kim, S. Y., Boutin, J. M., Glass, C. K. & Rosenfeld, M. G. (1991) Cell 67, 1251–1266.
38. Kliewer, S. A., Umesono, K., Mangelsdorf, D. J. & Evans, R. M. (1992) Nature (London) 355,446–449.
39. Falcone, M., Miyamoto, T., Fierro-Renoy, F., Macchia, E. & DeGroot, L. J. (1992)
Endocrinology 131, 2419–2429.
40. White, P. & Dauncey, M. J. (1999) J. Mol. Endocrinol. 23, 241–254.
41. Ness, G. C., Pendelton, L. C. & Zhao, Z. (1994) Biochim. Biophys. Acta. 1214, 229–233. 42. Ness, G. C. & Lopez, D. (1995) Arch. Biochem. Biophys. 323, 404–408.
43. Zilz, N. D., Murray, M. B. & Towle, H. C. (1990) J. Biol. Chem. 265, 8136–8143. 44. Dozin, B., Magnuson, M. A. & Nikodem, V. M. (1986) J. Biol. Chem. 261, 10290–10292. 45. Go¨the, S., Wang, Z., Ng, L., Kindblom, J. M., Campos Barros, A., Ohlsson, C., Vennstro¨m,
B. & Forrest, D. (1999) Genes Dev. 13, 1329–1341.
46. Refetoff, S., Weiss, R. E. & Usala, S. J. (1993) Endocr. Rev. 14, 348–399. 47. Forrest, D. & Vennstro¨m, B. (2000) Thyroid 10, 41–52.
‡‡Gauthier, K., Plateroti, M., Chassande, O. & Samarut, J. (2000) in Program, (The Endocrine Society, Toronto), p. 21 (abstr.).
